Before and after, with nothing edited out
This page will fill up slowly, because we publish measurements, not testimonials. Here is what a case looks like when it lands here, and what exists today.
What a case looks like here
Answer snapshot
The actual AI answer, before and after, quoted verbatim, so you can judge the change yourself instead of taking our word for it.
Matched-cohort time series
Citation frequency over a fixed query set: the same query × engine pairs, first 14 days versus the latest 14 days. No swapping queries mid-stream to flatter the trend.
Confidence label + confounders
Every figure carries a high / medium / low confidence label and lists the confounders: what else could plausibly explain the movement.
Ourselves
Our own product brand, measured with the exact protocol we sell.
Before offering measurement to anyone else, we pointed it at ourselves: a fixed buyer-question set, three runs per query, every engine we support. That baseline is now our starting line.
Six own-brand experiments, each with control queries, are starting shortly. As they complete, the full before/after data will appear on this page: answer snapshots, matched-cohort time series, confidence labels, dips included.
Where client cases will appear
We don’t have client cases to show yet. When they arrive, three rules apply to every single one:
- Published anonymized: industry and size tier, nothing more.
- Published with the client’s consent, every time.
- Never cherry-picked: the time series will include the weeks the numbers dipped.
Every after starts with a before.
Get your baseline measured now, so there is something to compare against later.
Run free scan →